Skip to main content

Full text of "Comparative fire resistance of roofing materials."

See other formats

% "6 


Fire Resistance of 
Roofing: cMaterials 

Published by 

Shingle Branch, West Coast Lumbermen's Association 

Seattle, Wash. 


In fairness to one of America's greatest indus- 
tries — an industry that for nearly a century has 
housed the nation — we ask those interested to 
give careful consideration to the report which 
appears herewith. 

This report is that of a fire test conducted by 
the Forestry Department of the University of 
Washington to ascertain the fire-resistant quali- 
ties of fifteen different roofs. 

The test was conducted by Professor Bror L. 
Grondal, and took place May 3, 1917. 

[1 ] 

Photographic Record of the Test with 
Explanation and Details 

^Sixteen different roofs were built according to standard building requirements, con- 
sisting of twelve of the leading advertised brands of prepared or patent roofings, also two 
treated and one untreated Red Cedar shingle roofs. 


The above is a photograph showing details of the standard fire brand used on all of 
the roofs. The group below shows how this fire brand was placed on the different roofs 
before ignition. 

' (>»;/v fifteen roofs iccrc tisid iti the test. 



h — ^' 1 

k: '^' ^^ ^" 


,^'' . 

k "^^^ »J 

k '^ " 

u- -^ " -. 



7 — ■ 

. - ..-....- .. , 


— ^ 





^ ^=-7^ 






— ' 





; 1 










Illustration A 

'S^r7c^c^r-<^ ///-(S- z5>«^^^ 

Tlie (liai:::rams on this page show the man- 
ner in whicli each section of the brand was 
selected. To avoid the shghtest variation be- 
tween any of the ^sixteen fire brands a piece 
of air-dried fir 46 inclies long- and 2 inches by 
2 inches cross section was carefullv selected. 
This piece just accommodated in leni^ih the 
eleven pieces or sections required for the 
brands. See illustration '*A." Each cross 
section permitted sixteen pieces >/, inch ])\' ^A 
inch from each leno-th. One of these i)ieces, 
Yz inch by J/ inch, was used in the same 
identical place in the construction of each of 
the sixteen different brands, thus ])()sitivelv 
eliminating- anv variation. The small sections 
were identical in form and size, and the 
brands when comi)leted weii/hcd exactlv the 
same to a fraction of an ounce. 

The excelsior used was carefullv weii^hed 
and ])lace(l in each brand in exactlv the same 

Just before io-niting- each brand was treat- 
ed with 15 c. c. of kerosene oil. 

The brands were all ionited at the same 
time and burned uniformly. The structure of 
the brands is explained in this leno-thly detail 
to show that every variable was eliminated as 
the value of the comparison depended on their 

* Sixteen hnuids built, fifteen used. 

- v^' 















Illustration B 

3 I 


i ligh ijrailc, 3-ply, prepared roohiiij. liefore i^iiiiun. 


of after fire had burned itself out. 

*A I 

\o. 1 — Sh»»uinf» «l;4in;«i:t 'V -in t" •-hcallnnj: 


No. .V — iliy;h ;j;radL-. '■ ]'1_\, jirt-parfil fiM.liiiL;. l^li'it iL^iiilion, 

. ih-if 

\o. 2 — Above roof after fire had bunuMi iisdf uui. 

Awd JO teccMid* mi' 

i« ) 

rr.^ tiMfeHMCL' J!L ^ " 

^^o. ;> — High grade, '2-ply, pre])are(l routing. Before ignitioi 

Xo. ;> — Above roof after hre had burned itself out. 

No. 3 — Showing damage done to sheathing. 

f ' ' , 

1 1 


-. ^ ^ ' 

1 ' 

K^- ^ 


^ -- 


\ .. 

fer ■ f 


I ^ 


No. 4 — RED CEDAR SHI.\<iLi:S treated with fire-retardant 
solution (commercial treatment). Before ignition. 


No. 4 — Above roof after fire had burned itself out. 

No. 4 — ^Showing >iic;minig not even scorched. 


Nos. 1, 2, o and 4 Roofs, .") minutes and *]<> seconds after ignition. 


. s-2 


I li-'ll 


\",, I, — llij^h iL^T'i'^*^' ;»<Uciit -]iiiiL;lc^. llfforc ii^nition. 

t attcr hr. 

\o. a — AUjvc rcK)f after tire iia<l lMinK<l itself (mt. 




^iKiwing dimi-i 

i<i i4icAihiii)^ . 

Showini^ <Uniagr done to Nhcathin^^. 

T— ~T 

1 ' 

Xo. 1 — Higli grade, 2-ply. prc]>ai'C(l roofing, llefore ignition. 

No. 8— RED CEDAR SHIXGLES treated with fire-rctardant. 

Xo. 'i — Al)ove roi.f after hre hail l);u'ned itself out. 




Xo. 8 — Above roof after fire had bnrncd itself out. 

Xo. N — Showino- sheathiTig not even -c irrl 

Xos. ."), (;. 7 and S Roofs, lo minutes after ignition. 




No. 9 — High grade, extra heavy, prepared roofing. Before ignition. 

^.; jfes^'f} 

^', ^^ 

No. 10 — High grade, 1-ply, prepared roofing. Before ignition. 

SJ-S «• 


Xo. I* — Above roof after fire had burnud il>elf out. 

Xo. it — Showing (hnnage done to sheathing. 

Xo. lU — Above ruuf after tire had burned itself out. 



-• — i 



l> 1 li I'' ^^i.;iiMi.dh^.Ji^^ 

Xo. Ill — Sliowiiij,' damage <l(jne to slieatliing. 

No.-. ;'. 1". 11 and l'.^ Kwi>. o 

•J n]lnutt^ ailer i^iiiition. 




.' '*^'i 














; • 


-W— r— — 1- 


/ • 



Xo. 11— RICl) CI':i)AR SillX(ifJ':S not pn.tcctrd ..r tnatd in 
any vvav. Ilctorc ii'tiilion. 

Xo. 12 — Hii^'-h i^radc patent >Iuni;lc>. Ilcforc i-uiin 





Xo. 11 — Ahovc ruof after tire lia<l htirrKtl itself out. 

No. 1*^ — .\l)()vc roof after fire had hiirned itself r)iit. 

No. U — Showing- that tire did not even >corch the third eonr> 
of ^h indies. 

In all these photographs the fact is clearly shown that 
Red Cedar Shingles did not support combustion nearly 
so readily as the patent roofings. 


//J .1'^ ^ 

- M^ 



Xo. 13 — High grade, prepared roofing. Before ignition. 


Xo. l;J — Al)()ve roof after fire had hnrncd itself out. 






Xt>. i;l — Showin*' daniai^e done to >heatliin<^ 

iJl^ u 




No. 14 — High grade, '^-ply, prepared roofing. Before ignition. 

\o. !4 — .\l)ove roof a tier hre had l)urned it-^elf o;U. 

\: - *^ ' 


Xo. li — Showing damage dune to >healhi: 

Xos. i;i. 14 and 1"» Kodfs. .; minute^ and ^l<» ^econ<ls after ignition. 


Xo. 1.-) — High i^radc, '^-ply. i)rci)arcd i-o(ihni>-. Before i.c^nition. 


Xo. lo — Above roof after fire had burned itself out. 

June loth, V)\7. 

StippleniciU tn Rc])urt of Ala\- ord on 
'*Ti:sTs ov RooFixc; AIatkrlm.s to Dktkrmixi-: 


The scries of tests made May .h'd wvvc 
(ki])Hcale(l June lilli. Tlie results (d* these 
check tests conformed closelx' with the orii^inal 
series. A pliotoi^raphic record of the second 
series is on hk at tlie L'niversitx' at Wash- 

i:r()R l. (;r( )\I)AL, 

^ I SSI stunt Frofcssor of Forestry, 
I 'ni\'crsit\' of // '(/.v7////<^'/<^//. 



III • r' fi 

No. lo — Showing- damage done to sheathing. 

Xos. 13, 14 and 1 •'» Roofs, 5 minutes after ignition. 

Test of Roofing Materials to Determine their 
Comparative Fire Resistance 

RROR L. GROXDAL, Jsst. Prof, of forestry, Uiikrrsify of irashiiigfon. 

Frecjitcnt assertions l)y manufacturers of roofino-s desio-ncd to rei)lace wooden shing-les that 
such "patent" roohn^^'s afford ])rotection from fire, while western red cedar sliingies presented 
an al)normally dano-erous fire hazard, have led many people to use such roofing's under the 
im])ression tliat they were "fire-proof." As fire records demonstrate rather conclusively that 
few fires orio-inate throu«-h the ig-nition of wooden shingles by sparks, the claims of some of 
these manufacturers were investigated by students at the College of Forestry of the University 
several years ago, and tlic discovery was made that any pitch or asphalt-saturated roofing paper, 
or "fire-resisting" roofing, would Imrn vigorously when lighted with an ordinary parlor match. 
It was found to be more difficult to ignite western red cedar shingles in a similar manner, and 
they (lid not burn so rapidly. The claims of the manufacturers of '*fire-resisting" or "patent" 
roofing were therefore viewed with some incredulity, but as exhaustive tests paralleling condi- 
tions which occur in general conflagrations had not been made, the results of these i)reliminar3^ 
exi)eriments were not given publicity. 

Early in the spring of 1017 Secretary H. P, Wyckoft* tendered tlie services of the Shingle 
Branch of the \\>st Coast Luml)ermen's Association in supplying material for a series of ]>rac- 
tical tests to determine the relative fire-resisting qualities of western red cedar shingles as com- 
l>ared with widely advertised sul)stitutes. As it was felt that such data, whether favorable toward 
western red cedar shingles or not, would be of great interest to the pu])lic, the ofi'er made l)y Mr. 
Wyckoft was immediately acce])ted bv the College of Forestry. After a practical method of 
testing, which eliminated all serious variables, had been devised and checked by preliminary tests, 
the tests described in the following i)aragraphs were carried out. 

Sixteen square roof sections, each having an area of sixteen square feet, w^ere prepared ac- 
cording to the methods recommended by the manufacturers of each type of roofing. The roof 
sections carrying ''patent" roofings were covered with 1x8 common shi])lap. The western red 
cedar sliingles were nailed to 1x3 common nailing strips. The sections were tested in lots of 
four. m{)unte(l on a common sui)])ort having a one-tlnrd ])itch, as illustrated in the photographs 
accom])anying this rei^ort. The following brands of r(»ofings were tested: 

No. 1 — \'ulcanite, 3 i)ly; Xo. 2 — Premium, 3 ply: Xo. 3 — Malthoid, 2 i)ly; Xo. 
4 — Kiln dried western red cedar shingles treated by immersion for one hour in 
Penetim firei)ro()fing lifjuid ; Xo. 5 — Paroid, 3 ply; XV>. 6 — Xe])onset shingles; X^o. 
7 — Amazon, 2 ])ly; Xo. H — Kiln dried western red cedar shingles treated with zinc 
borate by the l\ S. Forest Service at Madison, Wis. ; Xo. ^) — \Tilcanite, extra heavy; 
Xo. 10 — ^Rul>eroi(l, 1 i)ly; Xo. 11 — Kiln dried western red cedar shingles, untreated: 
Xo. 12 — Parrett Tylike shingles: Xo. 13 — Johns-Manville Asbestos roofing: Xo. 
1-t — Certainteed, 2 ]>ly: Xo. 15 — Paroid, 2 ply; Xo. 16 — Kiln dried western red ce- 
dar shingles treate(l with fireproof ])aint. ( X(». 16 was not tested, as the jiaint at 
the time of testing had not dried. ) 

After placing the roof sections upon the su|)port, a standard fireljrand built u]) with great 
care was attached to the roof in the center of the scjuare. The firel^rands were made from ma- 
terial cut from one close-grained Douglas fir plank. (See illustration i)age 3.) The weight of 
the wood and nails in each brand carried less than plus or minus 0.1 gram from an average 
weight of 114.5 grams. Kxactly 13 grams of spruce excelsior were packed into each brand, and 
to insure (juick and comj)lete ignition, 15 cubic centimeters of kerosene were si)rinkled over the 
excelsior in each brand immediately before they were ignited. To avoid the j)ossible accidental 
dis])lacemcnt of the brands during the test, they were hung f»n the roof sections by a small 
shingle nail tacked int(t the roof. The brands were placed as illustrated on j)age 2. 

The tests were carried out on May 3rd, 1917. The temi)erature during the testing of sections 
1. 2. 3 and 4 averaged about 70 degrees F. Wind was absent and the sun was shining. P*efore 
the tests on sections 5, (\ 7 and 8 were started the sky became overcast with clouds, obscuring 
the sun. while the temperature fell 8 degrees. Upon the C(»nii)letion of this lot. rain interrupted 
the testing, and further tests were discontinued for alxiut two hours, when the rain clouds dis- 
sipated. Sections ^>, 10, 11. 12, 13. 14 and 15, which had in the meantime been placed under 
cover, were tested while the temperature was about 65 degrees F. During these tests the sun 


was shining and wind was practically absent. It was noted that the first three sections tested 
were softer than those in the second rnn, and it is the opinion of the writer that had the tests 
])een carried out under a hot summer sun, the asphalt or pitch-saturated roofin^-s would ha\'e 
burned with even much greater vigor, while wooden shingles, containing no inrtammable bitu- 
minous compounds which soften with heat, would have been relatively unaffected. 

The four^ roof sections in each lot were ignited at the same instant by four attendants, and 
the time required for the roofing in each case to ignite, the time required for the roofing to 
l)urn through and the time required for the fire to burn out were carefully noted, with the fol- 
lowing results: 


Tituc required Time required 
Roof for roof for roofin<^ to 

No. to i^^nite burn through 


Time required Depth of ehar. Area of 
for fire to burn in sheatliini^. charriir^, 
itself out Inches Inches 



1 minute 

4 minutes 

10 minutes 




1 minute 

2 min. 20 sec. 

10 minutes 




1 minute 

3 min. 40 sec. 

18 min. 30 sec. 



Burned through sheath- 
ing in 12 min. 30 sec. 


1 min. 30 sec. 

Did not burn 

8 min. 30 sec. 





30 seconds 

3 minutes 

IS minutes 




35 seconds 

Did not burn 

15 min. 30 sec. 




30 seconds 

3 min. 30 sec. 

14 minutes 




2 min. 30 sec. 
began charrins^ 

Did not burn 

10 minutes 




Shingles charred, but 
(Hd not burn with 

40 seconds 

4 min. 30 sec. 13 minutes 



40 seconds 

4 minutes 

13 min. 30 sec. 



45 seconds 

Did not burn 

10 minutes 

Sb'i;lit trace 


30 seconds 

4 minutes 

14 min. 30 sec. 




45 seconds 

5 minutes 

9 minutes 




30 seconds 

2 minutes 

28 min. 30 sec. 



Sheathing practically 


30 seconds 

2 minutes 

10 min. 30 sec. 



The results as shown in tabulated form above show that the fire-resistance of western red 
cedar roofings is greater than any of the ''])atent'* roofings included in the test. Se\eral of the 
^'patent" roofings burned so vigorously that it is felt that they cannot be regarded as reasonablv 
safe fire hazards. All of these roofings, with the exception of No. 6, quickly burned througli l<> 
the sheathing. Roofing No, 6, however, burned vigorousl}-. None of the western red cedar 
shingle sections, including the kiln dried untreated shingles, Imrned through to the nailing strips. 
It became evident during the test that as the western red cedar shingles contained no asphalt 
or pitch, they did not support combustion long enough for the fire to spread, due to the low heat 
conductivity of the wood. Large volumes of infiammal)le gas distilled from the asphalt or pitch- 
saturated roofings and escaped through the seams in the sheathing below. In the case of a 
dwelling covered with such roofing and exposed to a conflagration it seems evident that the ac- 
cumulation of explosive gases in the attic would tremendously increase the fire hazard, for the 
flash of flame following the ignition of these gases would possibly ignite the framework of the 
entire roof. 


1. Roofings manufactured from paper containing rags, bark and similar combustible ma- 
terial should not be classed as fire-resisting roofings, for they readily support combustion, 

2. Roofings saturated with asphalt or pitch burn very readily, and even though the i)aper 
base may be of non-combustible materials, fire is readily communicated to the sheathing below. 
Such roofings should therefore not be classed as fire-resisting. 

3. Untreated western red cedar shingles constitute a relatively low fire hazard, and by the 
application of fireproofing compounds they can effectively be rendered fire-resisting. 



The following letters explain more clearly than 

we can why Birmingham repealed its 

Anti-Shingle Ordinance 


425-432 Henry Bldg, 
Seattle, A\'ash. 

Gcjifhviioi: Birmingham, Ala., April 15, 1917. 

Your letter of the 7th inst. received. 

For some time the City of Birmingham had on its statute hooks an anti-woodcn-shingle 
law. Last spring it hecame necessary for me to roof my house and I hegan an investigation of 
the various composition roofing materials re(|uired ]}y our city ordinance. I found that even the 
highest priced composition roofings were very unsatisfactory. I immediately began to agitate 
the repeal of the anti-wooden-shingle law and aroused great interest. Finally we induced the 
Commission to pass an ordinance i)ernntting the use of wooden shingles treated with certain 
kinds of ])aint. This law is now in effect in this city. 

Tn addition to a limited ]>ersonal ex])erience with composition roofing T have made a thor- 
ough investigation to determine its durability. fire-|>r(jof riualities, cost of jmtting it on, and 
cost ])er thousand shingles as com])are(l to tlie cost of wooden slhngles. Of course, I have 
not investigated ])rices excei)t in the local market. I find that a good grade of wooden shingle 
roof will last at least four times as long as even the ])est C(nnposition roof. When treated 
with a fire-proof paint the wooden shingle is as near fire-])roof as the com])osition roofing, 
es])ecially after the com])osition roofing has been on for some months and has been exjjosed 
to our \\arm sun. ^^*e find that here in the South the heat usually causes the coating of gravel 
to loosen and run off, lea\ing a dry ])aper \-ery inllammable. It costs more to lay a composi- 
tion ronf tban a wooden sliingle roof because of tbe necessity of doul)k' decking. And the cost of 
a good grade of wooden shingle is much less tban that of the composition shingle which is 
placed u])on the market as a reasonably good material. B.ut as I stated above, actual experi- 
ence has demonstrated tbat even the best composition roof will not last more than one fourth 
as long as the \\(»oden shingle roof. 

Scores of peo])le \\ho ba\'e been \ictims of com])osition roofing testified, at the x'arious 
hearings held by the City Commission (»f liirmingham for the jmrpose of investigating the 
relative merits of the two kinds of roofing, tr) tbe general unsatisfactoriness of the composition 
shingle. The evidence was conclusive on the pro])ositions that it is \ery difiicull to lay com])o- 
sition roofings so that tbey are weather-])roof even when first put on. that they have a very 
short life under any condition, that the gravel coating soon runs off, leax'ing a dry inflamnial)le 
])aj)er, and that they are generally unsatisfactftry. liotb sick's were well rei)resente<l at tbe bear- 
ings Init tbe evidence was so strong against tlie compositirm roofing and actual experience had 
l>roven it so inferior to the wooden shingle that tbe overwhehning verdict of the investigation 
was in condemnation of the com])osition roof. 

T have no objection to ])eing (juoted on tbe subject. I am ^o thoroughly convinced tbat laws, 
which com])el i>eople to use comi>osition roofs instead of tbe superior wooden shingle, work a 
grave injustice on the i)roi)erty owner, that I am ah\av< L^lad r>f an o])]Kjrtiinit\- to niter n word 
of protest against them. 

I am, \'ery trulv \ours, 
Mrs.W SAI. b.B. (Sif^ncd) MRS. \V. S. MORROW. 

I Ml 




George ]>. Ward, President 
James Weather ly J. D. Truss 

Arlie Barber J. R, Hornady 

C. li. r.lovd. Secretary 

I\IR. H. P. WYCKOFF, Seeretary, 
Henry I'nildin,^', 

Seattle, Washington. 

Dear Sir: 

Nov. C\ 1^M6. 

Replying- to your letter of October 30, I beg to enclose liere\\ith 
a copy of the ordinance adopted by the City Commission i)erniitting 
the use of wooden shingles within tlie cor])orate limits. 

This ordinance was adopted upon tlic rcf|uest of a large majorit\' 
of the proi)erty owners of the city. 


\'ery trulx* }'f)urs, 

(Si-iieiO C n. TJJ )^'^, Secretary. 

Houston repealed its ordinance for very much the same reasons, and numer- 
ous other cities after making investigations and tests decided not to put their 
citizens to the expense that Houston and Birmingham did. 

[15 1 

In Conclusion 

\Miat arc the obvious conclusions to be drawn from such an astonishing revelation as the 
])rece(lino- pag-es bring- forth? Does it not insist that Red Cedar Shingies must not be legis- 
lated against without the gravest inii)lication of prejudice or ignorance? 

Does it not i)ro\-e that the moulchng of iniblic sentiment by misleading pul)licity is danger- 
ous and accusations of com])etitors utterly unfounded? 

]>ut you sav — "\Miat alx)ut conflagrations?" True, wooden shingles burn in conflagrations. 
So do brick, tile, slate, iiiefal, composition — ez'crythiug. Authentic reports and post-fire investi- 
gations with very few exce])tions all i)rove that conflagrations invariably start in factory or fire- 
proof Imilding zones, and when fanned l)y wind destroy everything. 

Take for instance, the disastrous fires of Chicago, Baltimore, Boston, New Orleans, San 
Francisco, liangor. 

But we are not dealing with conflagrations which destroy everything, but with the resi- 
dence risk — the indi\'idual home, and statistics do not show that the wooden shingle roof is a 
menace comi)ared with other roofings. Jllicn put to the test Red Cedar Sliiui^les prove their 
superiority. \qs, wooden shingles, treated or untreated, without ])aint or fire-retardant, resist 
fire belter than the most exi)ensive, to say nothing of the cheai)est brands of ])repare(l roofings. 

Tt is not the desire of the manufacturers of wooden shingles to l)elilt]e or ()])]X)se sane fire 
])rotection, 1)ut is it not a mistake, is it not working a financial hardshi]) on the builder, is it 
not economically wrong when searching for fire protection to com])el a community by legisla- 
tion to replace a sui)erior prochtct with an inferior product? 


TX^^\RIAr>L^' fess hazardous than fref^ared roofixgs. 

Respectfully subnutted, 


Henry Building, Seattle, Wash.